[turba] upgrade problems

[Howard Moscovitz]

My problem is solved, but I see other people with similar problems so 
I'm sending this to the list. Maybe it will help. I upgraded from 1.1 to 
  2.0.4.

I did run 1.2_to_2.0.sql which didn't do the job. It turns out that 
there are other fields that we not in old database. I looked at the file

turba/scripts/sql/turba_objects.mysql.sql and manually added the missing 
fields to my db. Somehow between 1.1 and 2.0.4 some of the database 
fields names were changed to not have cap letters. I was able to 
manually adapt those.

When that was accomplished (my db has all the right fields and the are 
spelled correctly) I could add new records, but not see the old ones. 
Turns out the owner_id field in my old records had the hostname 
appended, like user at system.com. I dumped the db to a file and ran awk 
program on it to strip off the '@system.com' stuff and loaded it back 
in. Now everything seems to work fine fine. My users can see their 
address books.

There is one small issue. Now my database has the same fields as the 
current version, but they are in different order. Does anyone know if 
this matters? It is possible to write SQL that assumes the fields are in 
a specific order, but that is not good programming practice. Still, I 
wouldn't want to get burned.

For example, one could write something like

INSERT INTO turba.objects VALUES (data1, data1, data2, data3)

This assumes the files are ordered in the database in a certain way. A 
better way to write this code would be

INSERT INTO turba.objects  (column1, column1, column3, column4) VALUES 
(data1, data2, data3, data4)

I could reorder the fields with some more awk programming, if I had to. 
It is work I'd rather avoid it possible. Does the turba development team 
review these kinds of coding issues?

Anyone have any thoughts on this?

Thanks much...

Howard